21.1 The meaning of the terms ‘quality control’ and ‘Quality Assurance (QA)’ often vary according to the context. . In practical terms, QA relates to the overall measures taken by the laboratory to regulate quality, whereas quality control describes the individual measures which relate to the quality of individual samples or batches of samples.
21.2 As part of their quality systems, and to monitor day-to-day and batch-to-batch analytical performance, laboratories must operate an appropriate level of internal quality control (QC) checks and participate wherever possible in appropriate proficiency testing schemes (external QC). The level and type of QC will depend on criticality, nature of the analysis, frequency of analysis, batch size, degree of automation, and test difficulty and reliability.
21.3 Internal QC: This may take a variety of forms including the use of: blanks;
measurement standards; spiked samples; blind samples; replicate analysis and QC samples. The use of control charts is recommended, particularly for monitoring QC control samples (Ref C20-22).
21.3.1 The level of QC adopted must be demonstrably sufficient to ensure the validity of the results. Different types of quality control may be used to monitor different types of variation within the process. QC samples, analysed at intervals in the sample batch will indicate drift in the system; use of various types of blank will indicate what are the contributions to the instrument besides those from the analyte; duplicate analyses give a check of repeatability, as do blind samples. 21.3.2 QC samples are typical samples which are sufficiently stable and available in
sufficient quantities as to be available for analysis over an extended period of time. Over this period the random variation in performance of the analytical process can be monitored by monitoring the analysed value of the QC sample, usually by plotting it on a control chart. As long as the QC sample value is
acceptable, it is likely that results from samples in the same batch as the QC sample can be taken as reliable. The acceptability of the value obtained with the QC sample should be verified as early as practicable in the analytical process so that in the event of system failure as little effort as possible has been wasted on unreliable sample analysis.
21.3.3 It is the responsibility of the analyst to set and justify an appropriate level of quality control, based on a risk assessment taking into account the reliability of the method, and the criticality of the work. It is widely accepted that for routine analysis, a level of internal QC of 5% has been identified as reasonable, i.e., 1 in every 20 samples analysed should be a QC sample. However, for robust routine methods with high sample throughput, a lower level of QC may be reasonable. For more complex procedures, a level of 20% is not unusual and on occasions even 50% may be required. For analyses performed infrequently, a full system validation should be performed on each occasion. This may typically involve the use of a reference material containing a certified or known concentration of analyte, followed by replicate analyses of the sample and spiked sample (a sample to which a known amount of the analyte has been deliberately added). Those analyses undertaken more frequently should be subject to systematic QC procedures incorporating the use of control charts and check samples.
21.4 Proficiency testing (External QC): One of the best ways for an analytical laboratory to
monitor its performance against both its own requirements and the norm of other laboratories, is to participate regularly in proficiency testing schemes (Refer C7). Proficiency testing helps to highlight not only repeatability and reproducibility performance between laboratories, but also systematic errors, i.e. bias. Proficiency testing and other types of intercomparisons are accepted as being an important means of monitoring quality at national and international levels.
21.5 Accreditation bodies also recognize the benefit of these schemes as objective evidence of competence of the laboratory and of the effectiveness of the assessment process itself. Where possible, laboratories should select Proficiency Testing schemes which operate according to good international practice (Refer C7) and have transparent evidence of quality, eg. by accreditation or other peer review (Refer B16). Accredited laboratories are normally required to participate in proficiency testing, (where suitable schemes exist), as an integral part of their QA protocols. It is important to monitor proficiency testing results as a means of checking performance and to take corrective action as necessary.
22.
COMPUTERS AND COMPUTER CONTROLLED SYSTEMS
22.1 In chemical testing laboratories, computers have a wide variety of uses, including:
• control of critical environmental conditions;
• monitoring and control of inventories;
• calibration and maintenance schedules;
• stock control of reagents and measurement standards;
• design and performance of statistical experiments;
• control chart generation;
• monitoring of test procedures;
• control of automated instrumentation;
• capture, storage, retrieval, processing of data, manually or automatically;
• matching of sample and library data;
• generation of test reports;
• word processing;
• communication.
22.2 Interfaces and cables provide physical connections between different parts of the computer or between different computers. It is important that interfaces and cables are chosen to suit the particular application since they can seriously affect speed and quality of data transfer.
22.3 The chemical testing environment creates particular hazards for the operation of computers and storage of computer media. Advice can usually be found in the operating manuals, however particular care should be taken to avoid damage due to chemical, microbiological or dust contamination, heat, damp, and magnetic fields.
22.4 Initial validation should verify as many aspects of a computer's operation as possible. Similar checks should be carried out if the computer's use is changed, or after maintenance, or revision of software. Where a computer is used to gather and process data associated with chemical testing, for validation of that function, it is usually sufficient to assume correct operation if the computer produces expected answers when input with known parameters. Computer programs performing calculations can be validated by comparison with manually generated results. It should be noted that some faults will occur only when a particular set of parameters is input. In chemical testing, suitable checks on the data gathering and handling functions could be made using a Certified Reference Material for the initial validation, with a secondary measurement standard such as a quality control material used for regular repeat checks. Any recommendations made by the manufacturer should be taken into consideration. The validation procedure used for a particular system and any data recorded during validation should be documented. It may be difficult to validate these systems in isolation from the analytical instrument producing the original signal. Usually the whole system is validated in one go, by using chemical measurement standards or reference materials. Such validation is normally acceptable. It is convenient to illustrate validation using examples of typical applications:
22.4.1 Word-processing packages are widely used in laboratories to generate a wide variety of documentation. The laboratory should ensure that the use of word processing packages is controlled sufficiently to prevent the production of unauthorised reports or other documents. In the most simple cases, where the computer acts as little more than an electronic typewriter, validation is achieved by manually checking hard copies. More sophisticated systems read and process data to automatically produce reports in predetermined formats. Such systems will require additional checks.
22.4.2 Microprocessor controlled instruments will normally have a self-checking routine which is activated when the instrument is switched-on, and will include the recognition and checking of all peripheral equipment. Often the software is
not accessible. Under most circumstances validation can be performed by testing the various aspects of instrument function using known parameters, e.g. by testing reference materials, physical or chemical measurement standards or quality control samples.
22.4.3 Data handling or processing systems, integration systems. Before it can be processed, the output from the analytical instrument will usually need to be converted to a digital signal using an analogue/digital converter. The digitised data is then translated into a recognisable signal (numbers, peaks, spectra according to the system) by the software algorithm. The algorithm makes various decisions (such as deciding where peaks start and finish, or whether a number should be rounded up or down) according to programmed instructions. The algorithm is a common source of unexpected performance and validation should test the logic behind the decisions made by the algorithm.
22.4.4 Computer controlled automated system. This may embrace one or more of the foregoing examples, operated either simultaneously or in controlled time sequence. Such systems will normally be validated by checking for satisfactory operation (including performance under extreme circumstances) and establishing the reliability of the system before it is allowed to run unattended. The validation should consist of a validation of individual components, plus an overall check on the dialogue between individual components and the controlling computer. An assessment should be made of the likely causes of system malfunction. One important consideration is that the computer, interfaces and connecting cabling have sufficient capacity for the required tasks. If any part of the system is overloaded, its operation will slow down and possibly data may be lost. This could have serious consequences where the operations include time sequenced routines. Where possible the controlling software should be tailored to identify and highlight any such malfunctions and tag associated data. The use of quality control samples and standards run at intervals in the sample batches should then be sufficient to monitor correct performance on a day-to-day basis. Calculation routines can be checked by testing with known parameter values. Electronic transfer of data should be checked to ensure that no corruption has occurred during transmission. This can be achieved on the computer by the use of ‘verification files’ but, wherever practical, the transmission should be backed-up by a hard copy of the data.
22.4.5 Laboratory Information Management Systems. LIMS systems are increasingly popular as a way of managing laboratory activities. A LIMS is a computer based system with software which allows the electronic collation, calculation and dissemination of data, often received directly from analytical instruments. It incorporates word-processing, database, spreadsheet, and data processing capabilities and can perform a variety of functions, including: sample registration and tracking; test assignment and allocation; worksheet generation; processing captured data; quality control; financial control; and report generation. The operation of the LIMS may be confined to the laboratory itself or it may form part of a company wide computer system. Information may be input manually or downloaded directly from analytical instrumentation or other electronic devices such as bar-code readers. Information can be output either electronically or as hard-copies. Electronic outputs could consist of raw or processed data written to other computers either within the organisation, or remote, perhaps transmitted via
a modem or electronic mail. Similarly the information could be downloaded to a disk. Where data crosses from one system to another there may be a risk of data corruption through system incompatibility or the need to reformat the information. A well designed system enables high levels of QA to be achieved, right from the point of sample entry to the production of the final report. Particular validation requirements include management of access to the various functions, and audit trails to catalogue alterations and file management. Where data is transmitted electronically it will be necessary to build in safety checks to guard against data corruption and unauthorised access.